It occurred to me that the inkscape icons are vector in nature (or at least based on vector originals)....
And that there is (in the new build different preset icon sizes....)...
But if some web browsers can support vector file formats ETC.....

It occurred to me why not a vector app itself, IE with respect to the icons themselves and in realtime....

I.E. it has obviously been determined that for different sized windows, measured in pixel size ETC (i.E in cases where where hi-res and big monitors obviouly have much larger windows) that larger icon sizes are required...
And also obviously when viewing on smaller displays where the windows are allot smaller in pixel measurements ETC...that smaller icons are needed....

So..... would it be possible to (having pre-set a minimum and maximum size for the icons..) then have them re-size in real time depending on the size of the window itself.........?

Wouldn't that be awesome....Is that not what real vector icons should do? Scale....
Then you could have an additional % of normal size adjustment to set the default icon size for your default window size (full screen on your monitor)...
Then if you re-sized the window (smaller) the icons would actively reduce as you changed the window size until the minimum icon size was reached.....

Is this not a revolutionary idea?
It seems to me that even though inkscape's icons are vector derived that they still have a way to go before they truly take full advantage of what a vector format is all about...scalability...

If vector images are inherently scalable why not have true vector icons....IE that actually scale with the window (with in limits).

This really would make the icons vector proper....

Yes, this would be of course possible, but GTK probably doesn't support fluid button sizes. Even if it did it would still require some hefty code changes. Currently the icons are rendered once into pixel buffers at startup. So, they aren't even rerendered on redraws.

Also the icons look best at their intended size and most people probably won't want bigger icons at bigger resolutions (today's resolutions, that is).

If someone needs the icons larger in Inkscape, then they probably need them lager in all applications. They can do this already by changing the OS's display DPI setting. It seems redundant for Inkscape to go to the effort of scaling it's interface for these situations. Also, thanks to the popularity of LCDs which are optimised for a single resolution, varying sized monitors tend to display at the same DPI--there are exceptions that I've seen, but in those cases you do tend to modify the OS's DPI.

Im not sure what you mean by changing the OS's dpi....do you mean the display settings?
As i never change my display settings with an LCD...as LCDs are only good at displaying their native resolution properly...
And i think its not so good to have your display settings anything but the native res on of your display hardware...
As you will only force your display drivers to "emulate" the alien resolution by transformations that can only mean image degradation....(if thats what you ment...i dont know...who would buy a monitor that touts a massive 2000x1700 res or what ever and then halve that in software?... it would be madness...

Also i think if inkscape were not inherently interested in exploring and promoting vector formats they would not have used them for the basis of their icons in the first place...So im not convinced everyone will agree with you there (I.E.that scalability is not necessary ETC)...someone thought it was a hot idea to have vector icons for a vector app...And why not...
but they basically did it in such a way that even though the source is a vector image they are practically rasterized on a display level.....everything is but with out scalability vector icons may as well be raster ones full-stop....

if "they" did not wish to take full advantage of what vector brings to the table then they could just as well go back to raster icons and leave it at that....
But someone was thinking progressively....

Having pre-set sizes although an improvement is still a throwback to the raster days.....
And since i have been informed by another poster that the current icons are actually generated when the program boots
Then any size icons could be generated just as well....
even though it is true any image will look best at its native size....the range or practical sizes for icons is such that the existing vector images would serve more than adequately all the practical range of icon sizes anyone could hope for....

Having vector icons that scale is taking true advantage of vectors inate nature and at the same time eliminating several raster legacy bottlenecks in the process...
I.E. the need for different pre-set icon sets....
Having to lose sight of the bottom icons on the\side toolbar when you have reduced your window size only sightly ETC....

The advantages are obvious and the visual impact would be impressive...
Is it practically worth while from a man hour perspective? who knows....
All it takes is one coder that thinks its a good idea and were off....
hopefully these exist....
P.S. i was taking a look at your web site.... your inkscape tute is well done...and your art is quite good also....

David Hewitt wrote:Im not sure what you mean by changing the OS's dpi....do you mean the display settings?As i never change my display settings with an LCD...as LCDs are only good at displaying their native resolution properly...And i think its not so good to have your display settings anything but the native res on of your display hardware...

The DPI setting in Windows display property is different to resolution (although forgivably confusing). One digital pixel still equates to 1 screen pixel, but it scales the GUI. Fonts, icons, buttons, etc get bigger whilst a photo will remain at its original size.

David Hewitt wrote:who would buy a monitor that touts a massive 2000x1700 res or what ever and then halve that in software?... it would be madness...

I've seen laptops with such high resolution that even I had difficulty reading the text, and I don't wear glasses. I had to increase the DPI so the owner could read text on it. But isn't this what you are arguing for in Inkscape? The ability to double the size of the GUI and effectively reduce the screen real-estate?

David Hewitt wrote:Also i think if inkscape were not inherently interested in exploring and promoting vector formats they would not have used them for the basis of their icons in the first place.

Vectors have more benefits than just scalability--infact it's a recognised weakness when it comes to icons. Vectors are great for designing icons because of it's editability, but when scaling an icon, the crispness that you normally appreciate in vector is lost as strokes and lines scale to fractions of a pixel. The Tango set of icons are also designed using SVG, but each sized icon is individually designed to ensure crispness and the appropriate detail is clear. As an example, this is the official Inkscape logo scaled to fit in a 16x16 icon and this is a heavily modified and optomised version specifically for 16x16 If you were to scale-up it would start to look better. If you were to scale-up it would look worse.

Ok, you pass....(well said)

All you have said is true and has merit...
however:

My scaling idea did take into account a minimum icon size where by this scaling effect would be practical...
(For this reason....)

But your point is not lost on me.... it is obvious that if this were implemented....
That vector icons resized 20% even if they still looked cool could still none the less be tweaked in a raster application to be made markedly superior.

This is obvious..

But with monitors getting ever bigger this difference in quality (between an actively scaled vector and a custom tweaked icon of the same size) may not be so much as to condemn the advantages that such a scaling scheme would have....
Especially as even the small icons (on a big monitor) are allot bigger in pixel scale then your example....
(that thing would be a full stop on a big monitor).

There is still a place for this... just not at the extreme scales in your example...(with-in limits)
And like anything there is a trade off......with-in reasonable limits this scaling would be that impressive that 9 out of 10 people would not even care that static icons only available in 3 sizes would look better under their own optimized conditions.

dynamic scaling would be a bad idea IMO, the suitable icon size for a given monitor is a function of the monitor not the window size. If Smaller icons were usable, I'd be using them...

As to why we rasterize them, it's mainly a speed and memory thing, storing and retrieving a bitmap is quicker than constantly rerendering a vector.
The reason we render on startup was to eat dogfood, it's a very obvious way to check the renderer is behaving.

render speed to is a very valid concern but you should not let the current state of average video reader capability dictate what you would consider valid for a FUTURE evelopment such as this....
There is always someone with a cutting edge video card and a nice new fast PC that will want to max out their rendering options...
And what is"maxing it" today is common place tomorrow...

Personally, even though i'm going in to bat for this idea (like you) i can't see it happening (and i knew that when i posted this)... but that does not mean i'm not going to put the idea forward...
as with open source if one of the top 10 guys likes it then it can be as good as done.

But its good to hear all the objections that people have especially when some of my ideas are not polished proposals...
But for what its worth anything like this is worth a look as an option for those who have the rendering grunt to pull such things off.....
And like i said i cant see it happening but personally i would like to see it happen and i think such things (even though admittedly a bit of gimmick now) would attract allot of attention and win inkscape more users....
And this will not always be boarder-line viable.....
Pixel densities will rise (as they already are doing) and all of a sudden everyone will be doing this and then you will have missed your chance to be pioneers.

David Hewitt wrote:render speed to is a very valid concern but you should not let the current state of average video reader capability dictate what you would consider valid for a FUTURE evelopment such as this....
There is always someone with a cutting edge video card and a nice new fast PC that will want to max out their rendering options...
And what is"maxing it" today is common place tomorrow...

Except its nothing to do with video card speeds cos we dont use hardware acceleration currently. and if we ignored what the current state of tech was then we'd end up with an unusable product.
At some point the blue sky nice to haves have to meet the real world.

David Hewitt wrote:Pixel densities will rise (as they already are doing) and all of a sudden everyone will be doing this and then you will have missed your chance to be pioneers.

But pixel density is a constant on any given monitor, like i said. So scaling with the window isnt whats needed. Scaling more intelligently based on desktop dpi is a different matter...

Simarilius wrote:Except its nothing to do with video card speeds cos we dont use hardware acceleration currently. and if we ignored what the current state of tech was then we'd end up with an unusable product.
At some point the blue sky nice to haves have to meet the real world.

Someone said there may be render issues..... well if done properly there would probably not be....hardware acceleration or not...
And this is not as far out technically as the blue sky is distant.... Its not that big an ask technically or practically if done with-in reasonable constraints...

I like it, you don't.... thats the bottom line....(i think)

David Hewitt wrote:Its not that big an ask technically or practically if done with-in reasonable constraints...

Be careful to not break the golden rule of discussing ideas, otherwise we'll be expecting you to provide the code

I would if i could... believe me...